This framing of portfolio construction as an oscillating learning system feels useful, especially the line that the portfolio moves in order to know.

The place I would press is custodianship of the oscillation.

If allocation becomes a method of inquiry, then the portfolio is not only producing knowledge. It is perturbing a field of lives, places, dependencies, institutions, ecologies, and futures. So the questions of amplitude, frequency, reversibility, and timing also become questions of consent, consequence, and return.

Who is included in deciding how strongly the system is moved? Who carries the cost of the probe? Who receives the learning? How does the knowledge generated by the oscillation return to the field that made it possible?

Otherwise, disciplined variation can still become sophisticated extraction of signal from exposed systems.

Perhaps a complexity portfolio is not only an oscillating learning system. It is a custodial learning system: moving in order to know, but remaining accountable to those moved by the movement.